Studying the Underlying Event at CDF and the LHC

1
Studying the Underlying Event at CDF and the LHC
Leading Jet vs Z-Boson
Perugia, Italy October 27. 2008
Rick Field Craig Group Deepak Kar
Rick Field University of Florida
Outline of Talk

• The Towards, Away, and Transverse regions
  of h-f space.

• Four Jet Topologies plus Drell-Yan.

• The transMAX and transMIN regions.

• The observables First look at average
  quantities. Then do distributions.

• Look at ltpTgt versus Nchg in min-bias and
  Drell-Yan.

• Show some extrapolations of Drell-Yan to the LHC.

CDF Run 2
CMS at the LHC
2
UEMB_at_CMS
Initial Group Members Rick Field (Florida) Darin
Acosta (Florida) Paolo Bartalini
(Florida) Albert De Roeck (CERN) Livio Fano'
(INFN/Perugia at CERN) Filippo Ambroglini
(INFN/Perugia at CERN) Khristian Kotov (UF
Student, Acosta)
See the talks by Filippo Ambroglini and Florian
Bechtel tomorrow morning!
PTDR Volume 2 Section 3.3.2

• Measure Min-Bias and the Underlying Event at CMS

• The plan involves two phases.
• Phase 1 would be to measure min-bias and the
  underlying event as soon as possible (when the
  luminosity is low), perhaps during commissioning.
  We would then tune the QCD Monte-Carlo models
  for all the other CMS analyses. Phase 1 would be
  a service to the rest of the collaboration. As
  the measurements become more reliable we would
  re-tune the QCD Monte-Carlo models if necessary
  and begin Phase 2.
• Phase 2 is physics and would include comparing
  the min-bias and underlying event measurements
  at the LHC with the measurements we have done
  (and are doing now) at CDF and then writing a
  physics publication.

Perugia, Italy, March 2006
UEMB_at_CMS
Florida-Perugia-CERN
University of Perugia
3
QCD Monte-Carlo ModelsHigh Transverse Momentum
Jets
Underlying Event

• Start with the perturbative 2-to-2 (or sometimes
  2-to-3) parton-parton scattering and add initial
  and final-state gluon radiation (in the leading
  log approximation or modified leading log
  approximation).

• The underlying event consists of the beam-beam
  remnants and from particles arising from soft or
  semi-soft multiple parton interactions (MPI).

The underlying event is an unavoidable
background to most collider observables and
having good understand of it leads to more
precise collider measurements!

• Of course the outgoing colored partons fragment
  into hadron jet and inevitably underlying
  event observables receive contributions from
  initial and final-state radiation.

4
QCD Monte-Carlo ModelsLepton-Pair Production
Underlying Event

• Start with the perturbative Drell-Yan muon pair
  production and add initial-state gluon radiation
  (in the leading log approximation or modified
  leading log approximation).

• The underlying event consists of the beam-beam
  remnants and from particles arising from soft or
  semi-soft multiple parton interactions (MPI).

• Of course the outgoing colored partons fragment
  into hadron jet and inevitably underlying
  event observables receive contributions from
  initial-state radiation.

5
Towards, Away, Transverse
Look at the charged particle density, the charged
PTsum density and the ETsum density in all 3
regions!
Df Correlations relative to the leading
jet Charged particles pT gt 0.5 GeV/c h lt
1 Calorimeter towers ET gt 0.1 GeV h lt 1
Transverse region is very sensitive to the
underlying event!
Z-Boson Direction

• Look at correlations in the azimuthal angle Df
  relative to the leading charged particle jet (h
  lt 1) or the leading calorimeter jet (h lt 2).
• Define Df lt 60o as Toward, 60o lt Df lt 120o
  as Transverse , and Df gt 120o as Away.
  Each of the three regions have area DhDf 2120o
  4p/3.

6
Event Topologies

• Leading Jet events correspond to the leading
  calorimeter jet (MidPoint R 0.7) in the region
  h lt 2 with no other conditions.

Leading Jet
subset

• Inclusive 2-Jet Back-to-Back events are
  selected to have at least two jets with Jet1 and
  Jet2 nearly back-to-back (Df12 gt 150o) with
  almost equal transverse energies
  (PT(jet2)/PT(jet1) gt 0.8) with no other
  conditions .

Inc2J Back-to-Back
subset
Exc2J Back-to-Back

• Exclusive 2-Jet Back-to-Back events are
  selected to have at least two jets with Jet1 and
  Jet2 nearly back-to-back (Df12 gt 150o) with
  almost equal transverse energies
  (PT(jet2)/PT(jet1) gt 0.8) and PT(jet3) lt 15
  GeV/c.

Charged Jet

• Leading ChgJet events correspond to the leading
  charged particle jet (R 0.7) in the region h
  lt 1 with no other conditions.

• Z-Boson events are Drell-Yan events with 70 lt
  M(lepton-pair) lt 110 GeV with no other conditions.

Z-Boson
7
transMAX transMIN
transMIN very sensitive to the beam-beam
remnants!
Z-Boson Direction
Area 4p/6

• Define the MAX and MIN transverse regions
  (transMAX and transMIN) on an event-by-event
  basis with MAX (MIN) having the largest
  (smallest) density. Each of the two transverse
  regions have an area in h-f space of 4p/6.

• The transMIN region is very sensitive to the
  beam-beam remnant and the soft multiple parton
  interaction components of the underlying event.

• The difference, transDIF (transMAX minus
  transMIN), is very sensitive to the hard
  scattering component of the underlying event
  (i.e. hard initial and final-state radiation).

• The overall transverse density is the average
  of the transMAX and transMIN densities.

8
Observables at theParticle and Detector Level
Leading Jet
Observable Particle Level Detector Level
dNchg/dhdf Number of charged particles per unit h-f (pT gt 0.5 GeV/c, h lt 1) Number of good charged tracks per unit h-f (pT gt 0.5 GeV/c, h lt 1)
dPTsum/dhdf Scalar pT sum of charged particles per unit h-f (pT gt 0.5 GeV/c, h lt 1) Scalar pT sum of good charged tracks per unit h-f (pT gt 0.5 GeV/c, h lt 1)
ltpTgt Average pT of charged particles (pT gt 0.5 GeV/c, h lt 1) Average pT of good charged tracks (pT gt 0.5 GeV/c, h lt 1)
PTmax Maximum pT charged particle (pT gt 0.5 GeV/c, h lt 1) Require Nchg 1 Maximum pT good charged tracks (pT gt 0.5 GeV/c, h lt 1) Require Nchg 1
dETsum/dhdf Scalar ET sum of all particles per unit h-f (all pT, h lt 1) Scalar ET sum of all calorimeter towers per unit h-f (ET gt 0.1 GeV, h lt 1)
PTsum/ETsum Scalar pT sum of charged particles (pT gt 0.5 GeV/c, h lt 1) divided by the scalar ET sum of all particles (all pT, h lt 1) Scalar pT sum of good charged tracks (pT gt 0.5 GeV/c, h lt 1) divided by the scalar ET sum of calorimeter towers (ET gt 0.1 GeV, h lt 1)
Back-to-Back
9
CDF Run 1 PT(Z)
Tune used by the CDF-EWK group!
PYTHIA 6.2 CTEQ5L
Parameter Tune A Tune AW
MSTP(81) 1 1
MSTP(82) 4 4
PARP(82) 2.0 GeV 2.0 GeV
PARP(83) 0.5 0.5
PARP(84) 0.4 0.4
PARP(85) 0.9 0.9
PARP(86) 0.95 0.95
PARP(89) 1.8 TeV 1.8 TeV
PARP(90) 0.25 0.25
PARP(62) 1.0 1.25
PARP(64) 1.0 0.2
PARP(67) 4.0 4.0
MSTP(91) 1 1
PARP(91) 1.0 2.1
PARP(93) 5.0 15.0
UE Parameters
ISR Parameters

• Shows the Run 1 Z-boson pT distribution (ltpT(Z)gt
  11.5 GeV/c) compared with PYTHIA Tune A
  (ltpT(Z)gt 9.7 GeV/c), and PYTHIA Tune AW
  (ltpT(Z)gt 11.7 GeV/c).

Effective Q cut-off, below which space-like
showers are not evolved.
Intrensic KT
The Q2 kT2 in as for space-like showers is
scaled by PARP(64)!
10
Jet-Jet Correlations (DØ)

• MidPoint Cone Algorithm (R 0.7, fmerge 0.5)
• L 150 pb-1 (Phys. Rev. Lett. 94 221801 (2005))
• Data/NLO agreement good. Data/HERWIG agreement
  good.
• Data/PYTHIA agreement good provided PARP(67)
  1.0?4.0 (i.e. like Tune A, best fit 2.5).

11
CDF Run 1 PT(Z)
PYTHIA 6.2 CTEQ5L
Parameter Tune DW Tune AW
MSTP(81) 1 1
MSTP(82) 4 4
PARP(82) 1.9 GeV 2.0 GeV
PARP(83) 0.5 0.5
PARP(84) 0.4 0.4
PARP(85) 1.0 0.9
PARP(86) 1.0 0.95
PARP(89) 1.8 TeV 1.8 TeV
PARP(90) 0.25 0.25
PARP(62) 1.25 1.25
PARP(64) 0.2 0.2
PARP(67) 2.5 4.0
MSTP(91) 1 1
PARP(91) 2.1 2.1
PARP(93) 15.0 15.0
UE Parameters
ISR Parameters

• Shows the Run 1 Z-boson pT distribution (ltpT(Z)gt
  11.5 GeV/c) compared with PYTHIA Tune DW, and
  HERWIG.

Tune DW uses D0s perfered value of PARP(67)!
Intrensic KT
Tune DW has a lower value of PARP(67) and
slightly more MPI!
12
PYTHIA 6.2 Tunes
All use LO as with L 192 MeV!
Parameter Tune AW Tune DW Tune D6
PDF CTEQ5L CTEQ5L CTEQ6L
MSTP(81) 1 1 1
MSTP(82) 4 4 4
PARP(82) 2.0 GeV 1.9 GeV 1.8 GeV
PARP(83) 0.5 0.5 0.5
PARP(84) 0.4 0.4 0.4
PARP(85) 0.9 1.0 1.0
PARP(86) 0.95 1.0 1.0
PARP(89) 1.8 TeV 1.8 TeV 1.8 TeV
PARP(90) 0.25 0.25 0.25
PARP(62) 1.25 1.25 1.25
PARP(64) 0.2 0.2 0.2
PARP(67) 4.0 2.5 2.5
MSTP(91) 1 1 1
PARP(91) 2.1 2.1 2.1
PARP(93) 15.0 15.0 15.0
UE Parameters
Uses CTEQ6L
Tune A energy dependence!
ISR Parameter
Intrinsic KT
13
PYTHIA 6.2 Tunes
These are old PYTHIA 6.2 tunes! See the
talks by Arthur Moraes and Hendrik Hoeth for
new tunes!
All use LO as with L 192 MeV!
Parameter Tune DWT Tune D6T ATLAS
PDF CTEQ5L CTEQ6L CTEQ5L
MSTP(81) 1 1 1
MSTP(82) 4 4 4
PARP(82) 1.9409 GeV 1.8387 GeV 1.8 GeV
PARP(83) 0.5 0.5 0.5
PARP(84) 0.4 0.4 0.5
PARP(85) 1.0 1.0 0.33
PARP(86) 1.0 1.0 0.66
PARP(89) 1.96 TeV 1.96 TeV 1.0 TeV
PARP(90) 0.16 0.16 0.16
PARP(62) 1.25 1.25 1.0
PARP(64) 0.2 0.2 1.0
PARP(67) 2.5 2.5 1.0
MSTP(91) 1 1 1
PARP(91) 2.1 2.1 1.0
PARP(93) 15.0 15.0 5.0
UE Parameters
ATLAS energy dependence!
ISR Parameter
Intrinsic KT
14
JIMMY at CDF
JIMMY was tuned to fit the energy density in the
transverse region for leading jet events!
JIMMY Runs with HERWIG and adds multiple parton
interactions!
PT(JIM) 2.5 GeV/c.
The Energy in the Underlying Event in High PT
Jet Production
The Drell-Yan JIMMY Tune PTJIM 3.6
GeV/c, JMRAD(73) 1.8 JMRAD(91) 1.8
JIMMY MPI J. M. Butterworth J. R. Forshaw M. H.
Seymour
PT(JIM) 3.25 GeV/c.
Transverse ltDensitiesgt vs PT(jet1)
15
Charged Particle Density
HERWIG JIMMY Tune (PTJIM 3.6)

• Data at 1.96 TeV on the density of charged
  particles, dN/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson and Leading Jet events as a
  function of the leading jet pT or PT(Z) for the
  toward, away, and transverse regions. The
  data are corrected to the particle level (with
  errors that include both the statistical error
  and the systematic uncertainty) and are compared
  with PYTHIA Tune AW and Tune A, respectively, at
  the particle level (i.e. generator level).

16
Charged PTsum Density

• Data at 1.96 TeV on the charged scalar PTsum
  density, dPT/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson and Leading Jet events as a
  function of the leading jet pT or PT(Z) for the
  toward, away, and transverse regions. The
  data are corrected to the particle level (with
  errors that include both the statistical error
  and the systematic uncertainty) and are compared
  with PYTHIA Tune AW and Tune A, respectively, at
  the particle level (i.e. generator level).

17
The TransMAX/MIN Regions

• Data at 1.96 TeV on the charged particle density,
  dN/dhdf, with pT gt 0.5 GeV/c and h lt 1 for
  Z-Boson and Leading Jet events as a function
  of PT(Z) or the leading jet pT for the
  transMAX, and transMIN regions. The data are
  corrected to the particle level (with errors that
  include both the statistical error and the
  systematic uncertainty) and are compared with
  PYTHIA Tune AW and Tune A, respectively, at the
  particle level (i.e. generator level).

• Data at 1.96 TeV on the density of charged
  particles, dN/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for leading jet events as a function of the
  leading jet pT and for Z-Boson events as a
  function of PT(Z) for TransDIF transMAX
  minus transMIN regions. The data are corrected
  to the particle level (with errors that include
  both the statistical error and the systematic
  uncertainty) and are compared with PYTHIA Tune A
  and HERWIG (without MPI) at the particle level
  (i.e. generator level).

18
The TransMAX/MIN Regions

• Data at 1.96 TeV on the charged scalar PTsum
  density, dPT/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson and Leading Jet events as a
  function of PT(Z) or the leading jet pT for the
  transMAX, and transMIN regions. The data are
  corrected to the particle level (with errors that
  include both the statistical error and the
  systematic uncertainty) and are compared with
  PYTHIA Tune AW and Tune A, respectively, at the
  particle level (i.e. generator level).

• Data at 1.96 TeV on the charged scalar PTsum
  density, dPT/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for leading jet events as a function of the
  leading jet pT and for Z-Boson events as a
  function of PT(Z) for TransDIF transMAX
  minus transMIN regions. The data are corrected
  to the particle level (with errors that include
  both the statistical error and the systematic
  uncertainty) and are compared with PYTHIA Tune A
  and HERWIG (without MPI) at the particle level
  (i.e. generator level).

19
Charged Particle ltpTgt

• Data at 1.96 TeV on the charged particle average
  pT, with pT gt 0.5 GeV/c and h lt 1 for the
  toward region for Z-Boson and the
  transverse region for Leading Jet events as a
  function of the leading jet pT or PT(Z). The
  data are corrected to the particle level (with
  errors that include both the statistical error
  and the systematic uncertainty) and are compared
  with PYTHIA Tune AW and Tune A, respectively, at
  the particle level (i.e. generator level). The
  Z-Boson data are also compared with PYTHIA Tune
  DW, the ATLAS tune, and HERWIG (without MPI)

20
Z-Boson Towards, Transverse, TransMIN
Charge Density

• Data at 1.96 TeV on the density of charged
  particles, dN/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson events as a function of PT(Z) for
  the toward and transverse regions. The data
  are corrected to the particle level (with errors
  that include both the statistical error and the
  systematic uncertainty) and are compared with
  PYTHIA Tune AW and HERWIG (without MPI) at the
  particle level (i.e. generator level).

21
Z-Boson Towards, Transverse, TransMIN
Charge Density

• Data at 1.96 TeV on the charged scalar PTsum
  density, dPT/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson events as a function of PT(Z) for
  the toward and transverse regions. The data
  are corrected to the particle level (with errors
  that include both the statistical error and the
  systematic uncertainty) and are compared with
  PYTHIA Tune AW and HERWIG (without MPI) at the
  particle level (i.e. generator level).

22
Z-Boson Towards Region
HW without MPI

• Data at 1.96 TeV on the density of charged
  particles, dN/dhdf, with pT gt 0.5 GeV/c and h lt
  1 for Z-Boson events as a function of PT(Z) for
  the toward region. The data are corrected to
  the particle level (with errors that include both
  the statistical error and the systematic
  uncertainty) and are compared with PYTHIA Tune AW
  and HERWIG (without MPI) at the particle level
  (i.e. generator level).

23
Z-Boson Towards Region
HW (without MPI) almost no change!

• Data at 1.96 TeV on the average pT of charged
  particles with pT gt 0.5 GeV/c and h lt 1 for
  Z-Boson events as a function of PT(Z) for the
  toward region. The data are corrected to the
  particle level (with errors that include both the
  statistical error and the systematic uncertainty)
  and are compared with PYTHIA Tune AW and HERWIG
  (without MPI) at the particle level (i.e.
  generator level).

24
Charged ltPTgt versus Nchg
See the talk by Peter Skands on Thursday morning!
The charged ltPTgt rises with Nchg!

• Shows the average transverse momentum of charged
  particles (hlt1, pTgt0.5 GeV) versus the number
  of charged particles, Nchg, at the detector level
  for the CDF Run 2 Min-Bias events.

25
Min-Bias Correlations
New
See the talk by Niccolo Moggi this afternoon!

• Data at 1.96 TeV on the average pT of charged
  particles versus the number of charged particles
  (pT gt 0.4 GeV/c, h lt 1) for min-bias
  collisions at CDF Run 2. The data are corrected
  to the particle level and are compared with
  PYTHIA Tune A at the particle level (i.e.
  generator level).

26
Average PT versus Nchg
New

• Data at 1.96 TeV on the average pT of charged
  particles versus the number of charged particles
  (pT gt 0.4 GeV/c, h lt 1) for min-bias
  collisions at CDF Run 2. The data are corrected
  to the particle leveland are compared with PYTHIA
  Tune A, Tune DW, and the ATLAS tune at the
  particle level (i.e. generator level).

• Particle level predictions for the average pT of
  charged particles versus the number of charged
  particles (pT gt 0.5 GeV/c, h lt 1, excluding the
  lepton-pair) for for Drell-Yan production (70 lt
  M(pair) lt 110 GeV) at CDF Run 2.

27
Average PT(Z) versus Nchg
New
No MPI!

• Predictions for the average PT(Z-Boson) versus
  the number of charged particles (pT gt 0.5 GeV/c,
  h lt 1, excluding the lepton-pair) for for
  Drell-Yan production (70 lt M(pair) lt 110 GeV) at
  CDF Run 2.

• Data on the average pT of charged particles
  versus the number of charged particles (pT gt 0.5
  GeV/c, h lt 1, excluding the lepton-pair) for
  for Drell-Yan production (70 lt M(pair) lt 110 GeV)
  at CDF Run 2. The data are corrected to the
  particle level and are compared with various
  Monte-Carlo tunes at the particle level (i.e.
  generator level).

28
Average PT versus Nchg
New
PT(Z) lt 10 GeV/c
No MPI!
Remarkably similar behavior! Perhaps indicating
that MPI playing an important role in both
processes.

• Predictions for the average pT of charged
  particles versus the number of charged particles
  (pT gt 0.5 GeV/c, h lt 1, excluding the
  lepton-pair) for for Drell-Yan production (70 lt
  M(pair) lt 110 GeV, PT(pair) lt 10 GeV/c) at CDF
  Run 2.

• Data the average pT of charged particles versus
  the number of charged particles (pT gt 0.5 GeV/c,
  h lt 1, excluding the lepton-pair) for for
  Drell-Yan production (70 lt M(pair) lt 110 GeV,
  PT(pair) lt 10 GeV/c) at CDF Run 2. The data are
  corrected to the particle level and are compared
  with various Monte-Carlo tunes at the particle
  level (i.e. generator level).

29
Summary

• It is important to produce a lot of plots
  (corrected to the particle level) so that the
  theorists can tune and improve the QCD
  Monte-Carlo models. If they improve the
  transverse region they might miss-up the
  toward region etc.. We need to show the whole
  story!

• We are making good progress in understanding and
  modeling the underlying event in jet production
  and in Drell-Yan. Tune A and Tune AW describe
  the data very well, although not perfect.
  However, we do not yet have a perfect fit to all
  the features of the CDF underlying event data!

CDF Run 2 publication. Should be out by the end
of the year!

• Looking at ltpTgt versus Nchg in Drell-Yan with 70
  lt Mpair) lt 110 GeV and PT(pair) lt 5 GeV is a good
  way to look at MPI and the color connections. The
  data show the correlations expected from MPI!

• There are over 128 plots to get blessed and
  then to published. So far we have only looked at
  average quantities. We plan to also produce
  distributions and flow plots.

CDF-QCD Data for Theory

• I plan to construct a CDF-QCD Data for Theory
  WEBsite with the blessed plots together with
  tables of the data points and errors so that
  people can have access to the results .